Impact of social management on reproductive, adrenal and behavioural activity in the cheetah (Acinonyx jubatus)
INTRODUCTION
Wild cheetah (Acinonyx jubatus) populations are in decline throughout their native range, with probably fewer than 15,000 individuals remaining in Africa (Kraus & Marker-Kraus, 1992; Caro, 1994; Nowell &
Jackson, 1996). The cheetah is well known for several of its fascinating biological features, including low amounts of genetic variation (O’Brien et al., 1983) and consistently high proportions of pleiomorphic sperma- tozoa (Wildt et al., 1983, 1987). Although these and other biological phenomena have been studied in free- living cheetahs, many explicit characterizations simply cannot be undertaken in situ for logistical reasons.
Therefore, ex situpopulations can serve as an important
resource for controlled studies into mechanisms that are likely to be relevant to both in situ and ex situ conditions.
Interestingly, cheetahs held ex situare well known to reproduce poorly (Lindburg et al., 1993; Marker-Kraus
& Grisham, 1993), unlike their counterparts in nature (Laurenson, Caro & Borner, 1992; Caro, 1994). In the case of the cheetah, we are interested in linking in situ scholarly knowledge to help produce a self-sustaining ex situ population. Such a strategy supports free-living cheetahs by (1) developing viable hedge populations as
‘ambassadors’ for conservation, (2) educating the pub- lic about the plight of wild counterparts and (3) establishing a ‘research resource’ for investigating phe- nomena that cannot be examined in individuals ranging in nature. The cause of poor reproduction in captivity has been studied extensively, and it is now apparent that there is no health, genetic, nutritional or other physio- logical difference between breeding and non-breeding
DOI:10.1017/S1367943002004043 Printed in the United Kingdom
Nadja C. Wielebnowski1, Karen Ziegler2, David E. Wildt1, John Lukas2and Janine L. Brown1
1 Conservation and Research Center, National Zoological Park, Smithsonian Institution, Front Royal, Virginia 22630, USA
2White Oak Conservation Center, Yulee, Florida 32097, USA
(Received 16 November 2001; resubmitted 16 April 2002; accepted 14 June 2002)
Abstract
Cheetahs (Acinonyx jubatus) held ex situcan provide an important resource for obtaining new bio- logical information that usually cannot be gleaned from free-living individuals. However, consistent captive propagation of the cheetah, a prerequisite for establishing a self-sustaining population, has not been accomplished so far. This study examined the effect of a husbandry regimen commonly used in ex situ facilities on female cheetahs. Although generally solitary in the wild, zoos frequently house cheetahs in pairs or groups. Using non-invasive hormone monitoring and quantitative behavioural observations, we studied the impact of such enforced social conditions on behaviour and ovarian/adrenal activity. Eight female cheetahs were evaluated for two consecutive 6-month periods, first while maintained in pairs and then as individuals. Subsequently four females were regrouped into two new pairs and monitored for another 6 months. Females in five of six pairings demonstrated pro- longed anoestrus and displayed agonistic behaviours. After pair separation all females rapidly resumed oestrous cyclicity. Females in the sixth pair continued cycling throughout the year while consistently displaying affiliative grooming and no agonistic behaviours. Faecal corticoid patterns varied signifi- cantly among individuals, but appeared unrelated to behavioural or ovarian hormone patterns. Thus, data appear to indicate that same-sex pair-maintenance of behaviourally incompatible female chee- tahs may lead to suppressed ovarian cyclicity. This suppression appears linked to agonistic behav- iours but not to any particular adrenal hormone excretion pattern. Results clearly demonstrate the value of applying knowledge about in situ social behaviour to ex situ management practices.
Conversely, however, non-invasive hormone monitoring conducted ex situ may help us to identify physiological phenomena of potential relevance for future in situstudies.
All correspondence to: Dr Nadja Wielebnowski, Chicago Zoological Society, Brookfield Zoo, 3300 Golf Road, Brookfield, IL 60513. Tel: (708) 485-0263 ext. 251; Fax: (708) 485-3140;
E-mail: [email protected].
cheetahs (Zoo Biology 12, Special Issue on Cheetahs, 1993; Wildt et al., 1993; Caro, 1994). Instead, inappro- priate husbandry and management techniques are regarded as prime factors contributing to low breeding success (Lindburg et al., 1993; Caro, 1994; Brown et al., 1996).
In the wild, female cheetahs are usually solitary except when accompanied by dependent young (Caro, 1994).
In contrast, most ex situfacilities keep females in pairs or groups since overt aggression is rarely observed. This husbandry strategy maximizes space utilization while providing perhaps a more interesting exhibit for the pub- lic. However, inappropriate social conditions are known to lead to social stress and adrenal activation in other species, which, in turn, may compromise the immune system and/or reproduction (Moberg, 1985; Von Holst, 1998; Mendoza, Capitanio & Mason, 2000; Jasnow et al., 2001). Also, crowding and encounters with even moderately aggressive conspecifics can lead to chronic stress as evidenced by elevated blood cortisol levels (Fraser & Broom, 1990; Mendoza, 1991; Von Holst, 1998; Jasnow et al., 2001). Jurke et al. (1997) found higher faecal corticoid concentrations in female cheetahs that were reproductively inactive compared to females that cycled regularly. Terio & Munson (2000) recently found that gastritis and other non-infectious diseases in zoo-held cheetahs may be correlated with elevated glu- cocorticoid excretion in faeces.
In an extensive survey of the North American ex situ cheetah population, it was determined that most adult females had normal reproductive tract anatomy and pitu- itary function (Wildt et al., 1993). Yet, more than 50%
of females appeared acyclic on the basis of observing inactive ovaries (via laparoscopy) and parallel, one-time measurements of low circulating estradiol. A follow-up study longitudinally assessed follicular (oestrogen) and luteal (progestogen) activity by means of faecal steroid analysis in 26 females for up to 24 months (Brown et al., 1996). Twenty-five per cent of female cheetahs failed to exhibit any ovarian activity, and, of the 75% that did, none cycled continuously. Rather, cyclicity was inter- rupted by prolonged periods of anoestrus, often months in duration. No obvious environmental or seasonal factor was associated with this sporadic reproductive activity.
However, most of the cheetahs evaluated in the latter two studies were housed in pairs or groups. We there- fore speculated that cohabitation might have a suppres- sive effect on reproductive function. In the current study we test the hypothesis that housing female cheetahs together suppresses ovarian activity and that this effect is reflected by parallel increases in agonistic behaviours and glucocorticoid excretion (as an index of adrenal acti- vation, or ‘stress’). A secondary priority was to test the utility of faecal corticoid excretion as an index of stress in an intensively managed, controlled environment.
Faecal corticoid monitoring has been touted as useful for assessing adrenal activity of animals in captivity (e.g.
Palme & Möstl, 1997; Schatz & Palme, 2001;
Wielebnowski et al., 2002) and in nature (e.g. Wasser et al., 1997; Wasser et al., 2000; Foley, Papageorge &
Wasser, 2001; Harper & Austad, 2001). Similarly, non- invasive monitoring of reproductive hormones has been identified as a valuable conservation tool for investigat- ing ecological parameters in free-ranging populations (Berger et al., 1999). Thus, we speculated that this study may provide new information on the potential of non- invasive hormone monitoring for future studies of wild cheetahs, including assessing the impact of ecological variables and human and environmental disturbance on health and reproduction.
METHODS
Animals and facilities
Study animals were eight captive-born, adult female cheetahs (mean ± SEM, 7.5 ± 2.6 years of age; range, 3.5–11.9 years) maintained at the White Oak Conservation Center (Yulee, Florida, USA), a private facility closed to the public. At study onset, the Center maintained a total of eight male and ten female cheetahs in large outdoor chainlink enclosures. For this investi- gation, single and paired females were maintained in enclosures of about 2000 m2 that were immediately adjacent or in visual proximity to female conspecifics.
All enclosures had a grass substrate and mature trees, were similar in design, and offered equal numbers and quality of hiding opportunities. The diet was a com- mercial, ground-horsemeat-based product (Nebraska Brand Canine Diet; Central Nebraska Packing Co., North Platte, Nebraska, USA) supplemented occasion- ally with venison and ox tails. Animals were fed once daily, and water was provided ad libitum.
Experimental design
Each female was housed separately for at least 2 weeks before study onset and then randomly assigned to a pair.
Once established, pairs were maintained for 6 months, then the females were separated and housed alone for 6 months. At the end of the second 6-month period, four of the females were regrouped to form two new pairs for another 6 months of monitoring. Thus, the entire study comprised six different pairs (Table 1). Two were mother–daughter pairs (pairs 2 and 3), and all others were unrelated. Enclosures housing study females were dispersed randomly throughout the cheetah facility.
None of the females (pairs or singletons) was completely isolated from other male and female conspecifics, and
Table 1.List of study females and pairings
Pairs and identification Studbook Months spent
numbers numbers as pair
Pair 1: CH-1 & CH-2 2124 & 508 July–January Pair 2: CH-3 & CH-4* 405 & 2983 July–January Pair 3: CH-5 & CH-6* 1902 & 2920 July–January Pair 4: CH-7 & CH-8 1907 & 1834 July–January Pair 5: CH-5 & CH-7 1902 & 1907 June–December Pair 6: CH-6 & CH-8 2920 & 1834 June–December
* Mother–daughter pairs
all females had visual, auditory and occasional olfactory contact with at least one adult male. However, no direct physical interaction occurred between males and females during the study.
Behavioural data collection
Quantitative behavioural data were collected twice weekly on each female in 30-minute observation sessions between 0700–1100 hours. During intervals of paired housing, data were collected on both females simultane- ously. Types and definitions of recorded behaviours were consistent with previously published ethograms (Sorensen, 1995; Wielebnowski & Brown, 1998).
Continuous time sampling was used for measuring behavioural event frequencies, and instantaneous or inter- val sampling (1-minute time intervals) was used to esti- mate time spent in any behavioural state. A check sheet, timer and stopwatch were used, and most observations were carried out by two of the authors (NCW and KZ) with occasional assistance by two other trained observers.
Inter-observer reliability was assessed during six obser- vation sessions in which the four observers recorded behaviours simultaneously and independently on the same female. Observer agreement, as measured by the Kendall Coefficient of Concordance (W) (Martin &
Bateson, 1993), was 96% (W = 85 to W = 100, P< 0.05).
Faecal sample collection and steroid analysis Daily faecal samples were collected from all females from July to July over 12 consecutive months and then from June through November in the following year for the two sets of re-paired females (Table 1). Green food colouring (~1–2 ml) or uncooked corn (2 tablespoons) and rice (1 tablespoon) were added to the diet of one of the paired females to identify individual faecal samples.
Faecal samples were collected in 50-ml polypropylene conical vials that were labelled with animal number and date and stored frozen at –20°C. Samples were shipped on dry ice to the Conservation and Research Center for analysis at the end of each month.
Oestradiol, progestogen and corticoid metabolites were extracted from samples as previously described (Brown et al., 1994, 1996; Graham & Brown, 1996).
Briefly, samples were lyophilized and pulverized, and about 0.2 g of well-mixed powder was boiled in 5 ml aqueous ethanol (90% v/v) for 20 minutes. After cen- trifuging at 500 g for 10 minutes, the supernatant was recovered and the pellet resuspended in 5 ml 90%
ethanol, vortexed for 1 minute and recentrifuged. Both supernatants were combined, taken to dryness and redis- solved in 1 ml methanol. Extracts were vortexed (1 minute), placed into an ultrasonic cleaner for 30 sec- onds to free particles adhering to the vessel wall and revortexed briefly (15 seconds). Extract dilutions were made in assay buffer (0.01 mol PO4 l–1, 0.14 mol NaCl l–1, 0.01% NaN3, pH 7.4) before analysis by radioim- munoassays (RIA) (1:40 for oestradiol, 1:800–1:8000 for progestogens, 1:20–1:100 for corticoid metabolites).
Extraction recovery of [3H] oestradiol, [14C] proges- terone, or [3H] cortisol (New England Nuclear, Boston, MA) added to faecal samples prior to extraction exceeded 85%. Faecal oestradiol, progestogen and cor- ticoid metabolites were quantified using RIAs validated for the cheetah. The oestradiol RIA uses an antibody raised in rabbits against oestradiol-17β 6-o-carboxy- methyloxime: bovine serum albumin (provided by Dr Sam Wasser, University of Washington, Seattle, Washington, USA) (Risler, Wasser & Sackett, 1987), a 3H-labelled oestradiol tracer (New England Nuclear, USA) and oestradiol standards (Brown et al., 1994, 1996). The antibody cross-reacts 100% with oestradiol- 17β, 77% with 6-keto-oestradiol, 2% with 17α-hydroxy- progesterone and oestrone and < 1% with oestriol, 5-androstene-3β-17β-diol, testosterone, cholesterol and cortisol. Assay sensitivity, based on 90% of maximum binding, was 5 pg per tube.
The progesterone RIA relies on a monoclonal antibody produced against 4-pregnen-11-ol-3, 20-dione hemisuc- cinate: BSA (331; provided by Dr Jan Roser, University of California, Davis, California, USA), an 125I-labelled progesterone tracer (ICN Biomedical Inc., Costa Mesa, California, USA) and progesterone standards (Brown et al., 1994, 1996). The antibody cross-reacts 100% with progesterone, 96% with 5α-pregnane-3β-ol-20-one, 36%
with 5α-pregnane-3α-ol-20-one, 15% with 5β-pregnane- 3β-ol-one, 15% with 17β-hydroxyprogesterone, 13%
with pregnenolone, 7% with 5β-pregnane-3α-ol-20-on, 5% with 5β-pregnane-3α,17α-diol, 20α-one and < 1%
with pregnanediol-3-glucuronide, androstenedione, testosterone, oestrone, oestradiol, oestriol, 21-hydroxy- progesterone, 20α-hydroxyprogesterone and cortisol.
Assay sensitivity was 7.5 pg per tube.
Concentrations of corticoid metabolites were quanti- fied using a double antibody 125I-corticosterone RIA (ICN Biomedicals Inc., Costa Mesa, California, USA) (Graham & Brown, 1996; Terio, Citino & Brown, 1999).
The antiserum had the following crossreactivities:
100% corticosterone; and < 1% with desoxycorticos- terone, testosterone, cortisol, aldosterone, progesterone, androstenedione, dihydrotestosterone, cholesterol, dehydroepiandrosterone sulfate, 11-desoxycortisol, dihydroprogesterone, oestrone, oestradiol, oestriol, preg- nenolone, hydroxypregnenolone, and hydroxyproges- terone. Fifty µl of faecal extract (diluted 1:10 in assay buffer) were added to 50 µl of steroid diluent (provided in the assay kit) for analysis. Assay sensitivity was 12.5 ng/ml.
Intra- and interassay coefficients of variation were 9.3% and 13.6% for oestradiol, 7.6% and 10.7% respec- tively for progestogen, and 7.8% and 12.1% for corti- costerone RIAs, respectively. All faecal data are expressed as per gram dry weight.
Statistical analysis
StatViewSE & Graphics (version 551+), EXCEL (ver- sion 5.0) and JMP (version 3.1) for Macintosh were used to analyze data.
Recorded behaviours were converted into rates. First, the observed behavioural frequencies were adjusted for the time the focal animal was out of sight by subtracting the number of observation intervals during which the animal could not be seen from the total number of obser- vation intervals (a total of 30 1-minute intervals per observation session). Then the behavioural frequencies were divided by the remaining number of observation
intervals. Based on these calculations, the number of resulting behavioural rates for each 6-month period was added and divided by the number of observation sessions during each housing treatment (n = 52 sessions/6 months).
To compare differences in behavioural rates for each female and for all females combined during the two housing treatments, a Wilcoxon matched-pairs rank test was used (Siegel & Castellan, 1988). Differences in behavioural rates among pairs were tested using a Kruskal–Wallis one-way nonparametric analysis of variance (Sokal & Rohlf, 1995).
Significant increases in faecal oestradiol and corticoid metabolite concentrations were determined by an itera- tive process in which high values exceeding the mean plus 1.5 standard deviations (SD) were excluded (Brown et al., 1996). Baseline values were those remaining after excluding all high values. Individual average baseline concentrations were calculated from these remaining values. A peak was considered the highest point in a group of significantly (> 1.5 SD) elevated values.
Overall average concentrations included all values for each female. Oestrous cycle length was calculated as the number of days between peaks in oestradiol concentra- tions (presumed to be associated with oestrus) for per- iods not exceeding 30 days (i.e., twice the estimated oestrus cycle duration based on other studies; Eaton &
Craig, 1973; Asa et al., 1992). Inter-oestradiol peak intervals of > 30 days were considered periods of anoestrus (Brown et al., 1996). Because cheetahs are primarily induced ovulators (Wildt et al., 1993; Brown et al., 1996), progestogen increases were not expected.
Thus, this analysis was done retrospectively to determine the incidence of any spontaneous (non-mating induced) ovulation (defined as an increase in progestogen metabo- lite concentrations that exceeded the mean plus 2 SD and remained elevated for at least 1 week). A Student’s t-test was used to compare baseline, peak and overall averages of faecal oestradiol and corticoid concentra- tions for each female when housed in pairs versus alone;
and also to compare hormone concentrations of domi- nant versus subordinate individuals. Prior to analysis, data were tested for normality using the Shapiro–Wilk Test and for homoscedasticity using the Levine Test to ensure that the application of parametric statistics was appropriate (Sokal & Rohlf, 1995). Average values are presented as mean ± standard error (SEM).
RESULTS
Behavioural observations
Comparing behavioural frequencies among all females before and after pair separation revealed significant dif- ferences in the incidence of grooming (P= 0.003), pac- ing (P = 0.04) and flehmen (P = 0.02) (Fig. 1(a–c)).
Grooming activity decreased (P < 0.05, Fig. 1(a)), whereas pacing (Fig. 1(b)) and flehmen (Fig. 1(c)) increased (P< 0.05) when females were housed in pairs.
In ten of 12 cases, cheetahs only paced when pair-housed Fig. 1. Comparison of average rates (mean ± SEM) of auto-
grooming (a), pacing (b) and flehmen (c) observed in females while housed as pairs versus singletons (n= 52 observation sessions for each housing condition). An asterisk denotes sig- nificant differences (P < 0.05) within a female. Females labelled ‘a’ or ‘b’ were monitored during two pairings: ‘a’ rep- resents data collected during the first pairing and ‘b’ repre- sents data collected during the second pairing.
(Fig. 1(b)). All but one pair exhibited more (P< 0.05) agonistic behaviours (growl, hiss, chase, strike at, bite, attack, fight) (Fig. 2) than affiliative behaviours (allo- groom, social rub, play; P< 0.001, Fig. 2). Affiliative behaviours only were observed in Pair 2 (CH-3 & CH- 4) (Fig. 2) which was a mother and daughter pair. This relationship was not necessarily responsible for the compatibility, as Pair 3 (CH-5 & CH-6) also consisted of a mother and daughter, yet showed no affiliative behaviours.
Faecal oestradiol concentrations
Faecal oestradiol concentrations ranged from 7 to 420 ng/g dry faecal mass. For all females combined, overall average faecal oestradiol concentrations ranged from 67.6 ± 1.9 to 108.0 ± 2.5 ng/g. Average baseline concentrations ranged from 41.2 ± 1.3 to 71.5 ± 2.4 ng/g.
Average peak concentrations ranged from 109.2 ± 15.5 to 269.2 ± 4.4 ng/g. The duration of the oestrous cycle was highly variable among and within individuals, and prolonged anoestrus (> 30 days in duration) was observed in six of eight females and in five of the six pairings (Fig. 3(a–c)). In these cases, anoestrus lasted 68 to 170 days (Fig. 3(a–c)). In contrast, females in Pair 2 (CH-3 & CH-4) cycled continuously (Fig. 4(a)). When acyclic periods were excluded, the average oestrous cycle of the cheetah (n = 138 cycles) was 11.7 ± 3.9 days (range, 6 to 23 days). Within these cycles, oestradiol (on average) increased for 3.8 ± 0.4 days (range, 1–18 days).
In six of the eight females (five of the six pairs), over- all average, baseline and peak oestradiol concentrations were lower (P< 0.001) during pair-housing than when cheetahs were solitary (Figs 3(a–c); 5(a)). This effect also was observed in the four females that were re-paired after individual housing. In each case, faecal oestradiol
excretion was lower and females exhibited prolonged periods of anoestrus while together (Fig. 5(a)). Only Pair 2 (CH-3 & CH-4), the pair that also demonstrated affiliative behaviours (Fig. 2), experienced no differ- ences (P> 0.05) in oestradiol excretory dynamics before and after separation (Figs 4(a); 5(a)).
Faecal progestogen concentrations
Average baseline faecal progestogen concentrations among all females ranged from 0.6 to 8.0 µg/g. Two iso- lated instances of what appeared to be non-mating induced ovulation (on the basis of distinctive elevations in excreted progestogen) occurred in CH-2 and CH-4 when housed as singletons. These suspected ovulations may have been stimulated by courtship behaviour displayed by males introduced into nearby enclosures (see Brown et al., 1996). On these two occasions, fae- cal progestogen concentrations remained elevated above baseline for 41 and 38 days respectively, with peak con- centrations of 448.9 µg/g and 130.5 µg/g respectively.
Faecal corticoid concentrations
Faecal corticoid concentrations ranged from 5 to 730 ng/g. Overall average concentrations ranged from 33.9
± 1.3 to 68.5 ± 1.8 ng/g. Average baseline concentra- tions ranged from 28.9 ± 1.74 to 59.2 ± 1.8 ng/g, whereas peak values ranged from 104.2 ± 2.6 to 289.5 ± 3.1 ng/g.
Faecal corticoid concentrations in females housed in pairs and as singletons are presented in Figs 4(b), 5(b) and 6(a–c). While housed as pairs, corticoid concentra- tions were higher (P< 0.05) in two females (CH-6, CH- 7) (Figs 5(b); 6(b,c)) and lower (P < 0.05) in four females (CH-2, CH-4, CH-5, CH-8) (Figs 4(b); 5(b);
6(a–c)) than when housed as singletons. No differences between the two housing treatments (P > 0.05) were found for two females (CH-1, CH-3) (Figs 4(b); 5(b);
6(a)) and for one of the re-paired females (CH-6) (Figs 5(b); 6(b)) during the first pairing. For the other three re-paired females (CH-5, CH-7, CH-8), responses were consistent because females that had higher or lower cor- ticoid concentrations during the first pairing also had higher or lower concentrations during the second pair- ing (P< 0.05) (Fig. 5(b)).
DISCUSSION
The findings of this study are significant in that they reaffirm the rather simple notion that zoos should, when- ever possible, house wildlife according to social condi- tions observed in nature. This is especially important if one of the objectives is the establishment of self-sus- taining ex situpopulations of rare species. In the case of the cheetah, it appears that pairing females, a social sit- uation generally not observed in the wild, had a signif- icant impact on both ovarian function and behaviour.
In short, most females housed together experienced ovarian suppression and displayed more behaviours indicative of potential distress (i.e. stereotypic pacing, Fig. 2. Comparison of average rates (mean ± SEM) of ago-
nistic and affiliative behaviours observed in females while housed as pairs (n= 52). Females labelled ‘a’ or ‘b’ were mon- itored during two pairings: ‘a’ represents data collected dur- ing the first pairing and ‘b’ represents data collected during the second pairing.
increased aggression). Although pairing cheetahs resulted in altered behaviours and reproductive steroid patterns in females, findings could not be correlated to faecal corticoid excretion, a potential index of physio- logical stress. Particularly interesting was that separat- ing females into individual enclosures rapidly reversed ovarian suppression with re-cycling beginning almost immediately. These findings also emphasized the impor- tance of the solitary nature of cheetahs in the wild. Taken one step further, these results indicate the need to main- tain sufficient habitat size to avoid constrictions that might lead to excessive female-to-female encounters that, in turn, may lead to reduced reproductive fitness.
An inappropriate social environment ex situis known to inhibit reproduction in other species (Kleiman, 1980;
Lindburg & Fitch-Snyder, 1994; Estep & Drewsbury, 1996); and more generally, overcrowding has been shown to reduce reproductive success in social as well as non-social species (Christian & LeMunyan, 1958;
Boyd, 1986; Von Holst, 1998). For example, in an exten- sive study of small felids, Mellen (1991) found group size to be negatively correlated with reproductive suc- cess. For this reason, Mellen suggested that the preferred husbandry strategy for small-sized, wild felid species (genus Felis) is single housing with occasional intro- ductions to potential mates.
Our findings on female cheetahs are important because zoos, in an effort to conserve space and create interesting exhibits, frequently house this species in same-sex pairs or small groups. Present data suggest that Fig. 3. Representative faecal oestradiol profiles of female cheetahs that showed prolonged anoestrus while housed as pairs.
Pairs: (a) CH-1 and CH-2; (b) CH-5 and CH-6; (c) CH-7 and CH-8.
this practice may be contributing to the overall poor reproductive success of cheetahs in captivity. Therefore, it is recommended that the Cheetah Species Survival Plan (Cheetah SSP) programme, which currently man- ages all cheetahs in accredited North American zoos, and other captive-propagation programmes incorporate this knowledge into its management protocols. According to our findings female cheetahs that are a high priority for mating should be maintained in solitary conditions to promote normal ovarian activity. It is interesting to note that a reversal in ovarian suppression occurred rapidly after pair separation, even when previously pair-housed females were only separated into adjacent enclosures.
However, the long-term effect of such close proximity in single-housing situations on actual successful mating and conception remains to be tested.
Furthermore, it is noteworthy that only the use of intensive quantitative behavioural observations allowed us to detect the subtle antagonisms and the lack of affil- iative interactions in pair-housed females. Frequently, Fig. 4. Faecal hormone profiles of female cheetahs (CH-3 and CH-4) that showed no anoestrus periods while housed as pairs:
(a) oestradiol concentrations, (b) corticoid concentrations.
Fig. 5. (Left) Comparison of overall average faecal hormone concentrations in female cheetahs housed as pairs versus sin- gletons: (a) oestradiol concentrations; (b) corticoid concen- trations. Females labelled with ‘a’ or ‘b’ were monitored during two pairings: ‘a’ represents data collected during the first pairing and ‘b’ represents data collected during the second pairing.
managers of captive cheetahs characterize minor ago- nistic behaviours as trivial or unimportant to reproduc- tive health. However, our findings confirm that even subtle behavioural differences may be detrimental, trans- lating into ovarian suppression. Interestingly, one pair showed neither agonistic interactions nor prolonged ovarian quiescence. Familiarity may diminish aggression between individuals (Randall, 1989; Mendoza, 1991) and thus the mother–daughter relationship here may have facilitated the continuous cyclicity and affiliative behaviours. None the less, because agonistic behaviours and ovarian suppression were observed in the other mother–daughter pair, such a relationship cannot always be anticipated between close relatives.
Glucocorticoid hormones have been implicated in mediating reproductive success in several social species.
The resulting response mechanism has been termed
‘social stress response’, or in some, more extreme cases,
‘psychological castration’ (Brown, 1978; Creel et al., 1997; Von Holst, 1998). Thus, social stress due to increased aggression has been hypothesized to constitute a mechanism for reproductive suppression (Packard et al., 1985; Sapolsky, 1987; Fraser & Broom, 1990;
Wingfield, Hegner & Lewis, 1991; Blanchard et al., 1995). Many studies have demonstrated that acute or chronic emotional arousal or distress can lead to adreno- cortical activation and increased glucocorticoid produc- tion (Mason & Brady, 1956; Mason, 1968; Ursin & Olff, 1993; Jasnow et al., 2001). Jurke et al. (1997) have found that two out of seven zoo-held female cheetahs had high average faecal corticoid concentrations and also were acyclic. These authors speculated that social stress and variations in adrenal function might be contributing to variations in reproductive activity. However, we found no apparent relationship between ovarian and adrenal activity. Faecal corticoid concentrations were higher dur- Fig. 6. Representative faecal corticoid profiles of female cheetahs that showed prolonged anoestrus and aggression while housed in pairs: (a) CH-1 and CH-2; (b) CH-5 and CH-6; (c) CH-7 and CH-8.
ing pair-housing in three cases and lower in six cases, and no difference was found in three cases. Interestingly, there was evidence that the same individual may respond similarly to the pairing manipulation. For example, when CH-5 and CH-8 were re-paired, corticoid concentrations again decreased, whereas levels for CH-7 increased, as during the first pairing. However, the fourth female, CH- 6, experienced no difference in corticoid concentrations during the first pairing, but had higher corticoids during the second pairing. These results highlighted the indi- vidualistic nature of the stress response rather than any consistent trend among all females.
The disparate results from the faecal corticoid analy- sis also emphasized the challenges associated with defin- ing and measuring stress. Others have pointed out that, because hypothalamo–pituitary–adrenal function is modulated by multiple behavioural and physiological variables, an inconsistent relationship between social stress and adrenal activity may not be surprising (Saltzman et al., 1998; Wingfield & Ramenofsky, 1999).
Individuals of the same species, sex and age may differ markedly in response to the same environmental stimu- lus, as their behaviour may be based on individual ex- periences with aversive stimuli (Miller, 1980).
Therefore, measuring glucocorticoid concentrations alone may not always represent the best indicator of an individual’s stress response, and examination of other indicators is advised.
Quantitative behavioural observations conducted in this study identified some behaviours potentially associ- ated with stress. For example, stereotypic pacing, an indicator of suboptimal welfare for many captive species (Broom, 1983; Wiepkema, 1983; Broom & Johnson, 1993), was observed in all pair-housed females. Six of eight females only paced when in the same enclosure with another female. Furthermore, of the two females that paced while housed alone, both experienced increased pacing frequency when in a pair. The inci- dence of auto-grooming also declined for most individ- uals, possibly owing to the increased level of pacing and enhanced vigilance expended on monitoring activities of the other female in the same enclosure. The increased incidence of flehmen in seven of 12 cases may also have been related to enclosure mates needing to monitor each other’s markings via urine sniffing (Ewer, 1973; Eaton, 1974). Animals housed alone rarely exhibit this behav- iour. Pheromonal cues have been implicated in mediat- ing reproductive suppression in some social species (Wasser & Barash, 1983; Barrett, Abbott & George, 1990). This effect seems unlikely in this study, because Pair 2 had equal olfactory exposure, yet experienced no ovarian suppression. However, it is interesting to note that Pair 2 did not exhibit flehmen behaviour.
In summary, ovarian suppression occurs in female cheetahs that are maintained together in the same enclo- sure. The severity of the effect, however, may depend on pair composition and behavioural compatibility.
Although agonistic behaviour was observed in pairs that experienced ovarian suppression, increased glucocorti- coid concentrations were not associated with observed
aggression. This result may be due to (1) the idiosyn- cratic nature of the cheetah or the stress response itself, (2) marked individual variation with respect to behav- ioural and physiological characteristics and/or (3) different coping abilities and strategies of individuals.
Furthermore, the observed increase in stereotypic pacing and concurrent decline of auto-grooming in pair-housed females indicated some level of social stress not adequately reflected by the measurement of faecal glucocorticoids. Therefore, future studies should focus on observed ovarian suppression in combination with other potential measures of captivity stress, including immune function and general clinical health status.
Lastly, this study reconfirmed the value of non- invasive faecal hormone monitoring as a powerful tool for examining the impact of environmental/social factors and potential environmental disturbances on reproduc- tive function and adrenal activity in a wild species.
However, to better understand overall stress sensitivity and the effects of various environmental parameters on reproductive biology and ecology of cheetahs, it is essen- tial to carry out faecal steroid monitoring studies on wild populations. Recent studies on other species have already illustrated the feasibility of this method for mon- itoring the physiology/endocrinology of free-ranging populations (e.g. Creel et al., 1997; Wasser et al., 1997;
Berger et al., 1999; Cavigelli, 1999; Foley et al., 2001;
Harper & Austad, 2001). Measuring adrenal activity in wild cheetahs could be accomplished by closely moni- toring radio-collared individuals to collect fresh scat or through obtaining faecal swabs or samples during cap- ture for radio-collaring or relocation. One advantage is that the steroid hormone content of fresh faeces from captured or otherwise disturbed individuals represents the hormonal status of the individual 1–2 days before its capture, thus providing an accurate index of normative adrenal and reproductive hormone activity in truly wild individuals. Some of these studies are already in progress with colleagues working on wild cheetahs in Namibia.
Overall, the availability of faecal hormone monitoring provides important research opportunities for conserva- tion science, allowing us to obtain previously inaccessi- ble and highly relevant biological information for both in situ and ex situpopulations.
Acknowledgements
The authors thank: Rochelle Berman and Susan Hilton of the White Oak Conservation Center for assistance with behaviour data and sample collection; Kelly Hernandez and Cathy Rasler for sample shipment;
Alexandra Acco, Sonja Esch, Deja Gomes, Leslie Kordella, Ellen Muetstege, Shane Roberts and Brian Smith of the Conservation and Research Center for assisting with sample analysis and data entry; Astrid Bellem for technical support; and two anonymous reviewers for helpful comments. This study was sup- ported by the White Oak Conservation Center, the Scholarly Studies Program of the Smithsonian Institution and the Smithsonian Institution Women’s Committee.
REFERENCES
Asa, C. S., Junge, R. E., Bircher, S. J., Noble, G. A., Sarri, K. J.
& Plotka, E. D. (1992). Assessing reproductive cycles and preg- nancy in cheetahs (Acinonyx jubatus) by vaginal cytology.
Zoo Biol. 11: 139–151.
Barrett, J., Abbott, D. H. & George, L. M. (1990). Extension of reproductive suppression by pheromonal cues in subordinate female marmoset monkeys, Callithrix jacchus. J. Reprod. Fertil.
90: 411–418.
Berger, J., Ward, T., Roffe, T. & Monfort, S. L. (1999).
Conservation endocrinology: a noninvasive tool to understand relationships between carnivore colonization and ecological car- rying capacity. Conserv. Biol. 13: 980–989.
Blanchard, D. C., Spencer, R. L., Weiss, S. M., Blanchard, R. J., McEwen, B. & Sakai, R. R. (1995). Visible burrow systems as a model of chronic social stress: behavioral and neuroendocrine correlates. Psychoneuroendocr.20: 117–134.
Boyd, L. (1986). Behavior problems of equids in zoos. In The veterinary clinics of North American, equine practice, Vol. 2(3):
653–664. Crowell-Davis, S. L. & Houpt, K. A. (Eds).
Philadelphia: W. B. Saunders.
Broom, D. M. (1983). Stereotypies as animal welfare indicators.
In Indicators relevant to farm animal welfare: Current Topics in Veterinary, Medical Animal Science, 23: 81–87. Schmid, D.
(Ed.). The Hague: Martinus Nijhoff.
Broom, D. M. & Johnson, K. G. (1993). Stress and animal wel- fare. London: Chapman & Hall.
Brown, J. L. (1978). Avian communal breeding systems. Annu.
Rev. Ecol. Syst.9: 123–155.
Brown, J. L., Wasser, S. K., Wildt, D. E. & Graham, L. H. (1994).
Comparative aspects of steroid hormone metabolism and ovar- ian activity in felids, measured non-invasively in feces. Biol.
Reprod.51: 776–786.
Brown, J. L., Wildt, D. E., Wielebnowski, N., Goodrowe, K. L., Graham, L. H., Wells, S. & Howard, J. G. (1996). Reproductive activity in captive female cheetahs (Acinonyx jubatus) assessed by faecal steroids. J. Reprod. Fertil.106: 337–346.
Caro, T. M. (1994). Cheetahs of the Serengeti Plains: group liv- ing in an asocial species. Chicago: University of Chicago Press.
Cavigelli, S. A. (1999). Behavioural patterns associated with fae- cal cortisol levels in free-ranging female ring-tailed lemurs, Lemur catta. Anim. Behav. 57: 935–944.
Christian, J. J. & LeMunyan, C. D. (1958). Adverse effects of crowding on lactation and reproduction of mice and two gener- ations of their offspring. Endocrinology63: 517–529.
Creel, S. R., Creel, N. M., Mills, G. L. & Monfort, S. L. (1997).
Rank and reproduction in cooperatively breeding African wild dogs: behavioral and endocrine correlates. Behav. Ecol. 8:
298–306.
Eaton, R. L. (1974). The cheetah: biology, ecology and behavior of an endangered species. New York: Van Nostrand Reinhold.
Eaton, R. L. & Craig, S. J. (1973). Captive management and mat- ing behavior of the cheetah. In The world’s cats I: ecology and conservation: 217–254. Eaton, R. L. (Ed.). Winston, OR:
Wildlife Safari.
Estep, D. Q. & Dewsbury, D. A. (1996). Mammalian reproduc- tive behavior. In Wild mammals in captivity: 379–389. Kleiman, D. G., Allen, M. E., Thompson, K. V. & Lumpkin, S. (Eds).
Chicago: University of Chicago Press.
Ewer, R. F. (1973). The carnivores. Ithaca: Cornell University Press.
Foley, C. A. H., Papageorge, S. & Wasser, S. K. (2001).
Noninvasive stress and reproductive measures of social and eco- logical pressures in free-ranging African elephants. Conserv.
Biol. 15(4): 1134–1142.
Fraser, A. F. & Broom, D. M. (1990). Farm animal behaviour and welfare. Paris: Baillière Tindall.
Graham, L. H. & Brown, J. L. (1996). Cortisol metabolism in the
domestic cat and implications for developing a non-invasive measure of adrenocortical activity in non-domestic felids. Zoo Biol.15: 71–82.
Harper, J. M. & Austad, S. N. (2001). Effect of capture and season on fecal glucocorticoid levels in deer mice (Peromyscus manic- ulatus) and red-backed voles (Clethrionomys gapperi). Gen.
Comp. Endocrinol.123: 337–344.
Jasnow, A. M., Drazen, D. L., Huhman, K. L., Nelson, R. J. &
Demas, G. E. (2001). Acute and chronic social defeat suppresses humoral immunity of male Syrian hamsters (Mesocricetus aura- tus). Horm. Behav.40: 428–433.
Jurke, M. H., Czekala, N. M., Lindburg, D. G. & Millard, S. E.
(1997). Fecal corticoid metabolite measurement in the cheetahs (Acinonyx jubatus). Zoo Biol.16: 133–147.
Kleiman, D. (1980). The sociobiology of captive propagation. In Conservation biology: 243–261. Soulé, M. E. & Wilcox, B. A.
(Eds). Sunderland, MA: Sinauer Associates.
Kraus, D. & Marker-Kraus, L. (1992). Cheetah preservation fund report. Cat News 17: 12–14.
Laurenson, M. K., Caro, T. M. & Borner, M. (1992). Female cheetah reproduction. Natl. Geog. Res. Expl.8: 64–75.
Lindburg, D. G., Durrant, B. S., Millard, S. E. & Oosterhuis, J. E. (1993). Fertility assessment of cheetah males with poor quality semen. Zoo Biol.12: 97–103.
Lindburg, D. G. & Fitch-Snyder, H. (1994). Use of behavior to evaluate reproductive problems in captive mammals. Zoo Biol.
13: 433–455.
Marker-Kraus, L. & Grisham, J. (1993). Captive breeding of the cheetahs in North America: 1987–1991. Zoo Biol.12: 5–18.
Martin, P. & Bateson, P. (1993). Measuring behaviour: an intro- ductory guide. Cambridge: Cambridge University Press.
Mason, J. W. (1968). A review of psychoendocrine research on the pituitary–adrenal cortical system. Psychosom. Med. 30:
576–607.
Mason, J. W. & Brady, J. V. (1956). Plasma 17-hydroxycorti- costeroid changes related to reserpine effects on emotional behavior. Science124: 983–984.
Mellen, J. D. (1991). Factors influencing reproductive success in small captive exotic felids (Felis ssp.): a multiple regression analysis. Zoo Biol.10: 95–110.
Mendoza, S. P. (1991). Behavioural and physiological indices of social relationships: comparative studies of New World mon- keys. In Primate responses to environmental change:311–335.
Box, H. O. (Ed.). London: Chapman & Hall.
Mendoza, S. P., Capitanio, J. P. & Mason, W. A. (2000). Chronic social stress: studies in non-human primates. In The biology of animal stress: 227–247. Moberg, G. P. & Mench, J. A. (Eds).
Wallingford: CAB International.
Miller, S. M. (1980). When is a little information a dangerous thing? Coping with stressful events by monitoring versus blunt- ing. In Coping and health: 145–169. Levine, S. & Ursin, H.
(Eds). New York: Plenum Press.
Moberg, G. P. (1985). Influence of stress on reproduction: a mea- sure of well-being. In Animal stress:27–50. Moberg, G. P. (Ed.).
Bethesda: American Physiological Society.
Nowell, K. & Jackson, P. (Eds) (1996). Wild cats. Status Survey and Conservation Action Plan. Gland, Switzerland: IUCN/SSC Cat Specialist Group, IUCN.
O’Brien, S. J., Wildt, D. E., Goldman, D., Merrill, C. R. & Bush, M. (1983). The cheetah is depauperate in genetic variation.
Science 221: 459–62.
Packard, J. M., Seal, U. S., Mech, L. D. & Plotka, E. D. (1985).
Causes of reproductive failure in two family groups of wolves.
Z. Tierpsychol.68: 24–40.
Palme, R. & Möstl, E. (1997). Measurement of cortisol metabolites in faeces of sheep as a parameter of cortisol con- centration in blood. Int. J. Mammal. Biol. 62 (Suppl. II):
192–197.
Randall, J. A. (1989). Territorial defense interactions with neigh-
bors and strangers in banner-tailed kangaroo rats. J. Mammal.
70: 308–315.
Risler, L., Wasser, S. K. & Sackett, G. P. (1987). Measurement of excreted steroids in Macaca nemestrina. Am. J. Primatol.12:
91–100.
Saltzman, W., Schultz-Darken, N. J., Wegner, F. H., Wittwer, D. J. & Abbott, D. H. (1998). Suppression of cortisol levels in subordinate female marmosets: reproductive and social contri- butions. Horm. Behav.33: 58–74.
Sapolsky, R. M. (1987). Stress, social status and reproductive physiology in free-living baboons. In Psychobiology of repro- ductive behavior: an evolutionary perspective:292–322. Crews, D. (Ed.). Englewood Cliffs: Prentice Hall.
Schatz, S. & Palme, R. (2001). Measurement of faecal cortisol metabolites in cats and dogs: a noninvasive method to evalu- ate adrenocortical function. Vet. Res. Comm.25: 271–287.
Siegel, S. & Castellan, N.J. Jr. (1988). Nonparametric statistics for the behavioral sciences. New York: McGraw-Hill.
Sokal, R. R. & Rohlf, F. J. (1995). Biometry. New York: W.H.
Freeman.
Sorensen, C. (1995). The cheetah (Acinonyx jubatus) in captivity:
descriptions of estrus and fecal steroid fluctuations, activity pat- terns and solitary behavior. In International Studbook, Cheetah, 1993: section K10. Marker-Kraus, L. (Ed.). Washington, DC:
National Zoological Park, Smithsonian Institution.
Terio, K. S., Citino, S. B. & Brown, J. (1999). Fecal cortisol metabolite analysis for non-invasive monitoring of adrenocorti- cal function in the cheetah (Acinonyx jubatus). J. Zoo Wildl.
Med.30: 484–491.
Terio, K. S. & Munson, L. (2000). Gastritis in cheetahs and relat- edness to adrenal function. In Felid taxonomy advisory group action plan, 2000 report: 36. Pukazhenthi, B., Wildt, D. &
Mellen, J. (Eds). Wheeling, West Virginia: American Zoo and Aquarium Association.
Ursin, H. & Olff, M. (1993). The stress response. In Stress: from the synapse to the syndrome: 3–22. Stanford, S. C. & Salmon, P. (Eds). London: Academic Press.
Von Holst, D. (1998). The concept of stress and its relevance for animal behavior. In Stress and behavior: 1–132. Advances in the Study of Behavior 2.7. Møller, A. P., Milinski, M. & Slater, P. J. B. (Eds). London: Academic Press.
Wasser, S. K. & Barash, D. P. (1983). Reproductive suppression
among female mammals: implications for biomedicine and sexual selection theory. Quart. Rev. Biol.58: 513–538.
Wasser, S. K., Bevis, K., King, G. & Hanson, E. (1997).
Noninvasive physiological measures of disturbance in the north- ern spotted owl. Conserv. Biol. 11: 1019–1022.
Wasser, S. K., Hunt, K. E., Brown, J. L., Cooper, K., Crockett, C. M., Bechert, U., Millspaugh, J. J., Larson, S. & Monfort, S. L. (2000). A generalized fecal glucocorticoid assay for use in a diverse array of nondomestic mammalian and avian species.
Gen. Comp. End.120: 260–275.
Wielebnowski, N. & Brown, J. L. (1998). Behavioral correlates of physiological estrus in cheetahs. Zoo Biol.17: 193–209.
Wielebnowski, N. C., Fletchall, N., Carlstead, K., Busso, J. M. &
Brown, J. L. (2002). Non-invasive assessment of adrenal activ- ity associated with husbandry and behavioral factors in the North American clouded leopard population. Zoo Biol. 21: 77–98.
Wiepkema, P. R. (1983). On the significance of ethological cri- teria for the assessment of animal welfare. In Indicators Relevant to Farm Animal Welfare: Current Topics in Veterinary Medical Animal Science 23: 69–79. Schmid, D. (Ed.). The Hague, Netherlands, Martinus Nijhoff.
Wildt, D. E., Brown, J. L., Bush, M., Barone, M. A., Cooper, K.
A., Grisham, J. & Howard, J. G. (1993). Reproductive status of cheetahs (Acinonyx jubatus) in North American zoos: the ben- efits of physiological surveys for strategic planning. Zoo Biol.
12: 45–80.
Wildt, D. E., Bush, M., Howard, J. G., O’Brien, S. J., Meltzer, D., van Dyk, A., Ebedes, H. & Brand, D. J. (1983). Unique sem- inal quality in the South African cheetah and a comparative eval- uation in the domestic cat. Biol. Reprod. 29: 1019–1025.
Wildt, D. E., O’Brien, S. J., Howard, J. G., Caro, T. M., Roelke, M. E., Brown, J. L. & Bush, M. (1987). Similarity in ejaculate characteristics in captive versus free-ranging cheetahs of two subspecies. Biol. Reprod. 36: 351–60.
Wingfield, J. C., Hegner, R. E. & Lewis, D. M. (1991). Circulating levels of luteinizing hormone and steroid hormones in relation to social status in the cooperatively breeding white-crowned sparrow weaver, Plocepasser mahali. J. Zool. Lond.225: 43–58.
Wingfield, J. C. & Ramenofsky, M. (1999). Hormones and behav- ioral ecology of stress. In Stress Physiology in Animals: 1–51.
Balm, P. H. M. (Ed.). Sheffield, UK: Sheffield Academic Press.
